U.S. patent number 8,731,941 [Application Number 13/220,317] was granted by the patent office on 2014-05-20 for voice-activated measurement system.
This patent grant is currently assigned to Zanavox. The grantee listed for this patent is David Edward Newman. Invention is credited to David Edward Newman.
United States Patent |
8,731,941 |
Newman |
May 20, 2014 |
Voice-activated measurement system
Abstract
A voice-activated instrument performs a measurement and displays
the measured value when commanded by voice. The system also resets
under voice control. The measurement trigger is any single-syllable
command such as "Count" or "Go". The reset trigger is any
two-syllable command such as "Reset". Any type of momentary
measurement device may be controlled in this way, including time
interval measurements, event counting, length measuring, weighing,
and electronic metering measurements, and many others.
Inventors: |
Newman; David Edward (Temecula,
CA) |
Applicant: |
Name |
City |
State |
Country |
Type |
Newman; David Edward |
Temecula |
CA |
US |
|
|
Assignee: |
Zanavox (Temecula, CA)
|
Family
ID: |
47744891 |
Appl.
No.: |
13/220,317 |
Filed: |
August 29, 2011 |
Prior Publication Data
|
|
|
|
Document
Identifier |
Publication Date |
|
US 20130054246 A1 |
Feb 28, 2013 |
|
Current U.S.
Class: |
704/275;
177/25.12 |
Current CPC
Class: |
G01B
3/1061 (20130101); G10L 15/02 (20130101); G01B
3/205 (20130101); G01R 1/025 (20130101); G10L
2015/223 (20130101); G10L 2015/027 (20130101); G01B
3/1069 (20200101) |
Current International
Class: |
G10L
21/00 (20130101); G01G 19/34 (20060101) |
Field of
Search: |
;704/270,275
;702/173-175 ;177/25.11-25.14 |
References Cited
[Referenced By]
U.S. Patent Documents
Primary Examiner: Azad; Abul
Claims
What is claimed is:
1. A system comprising a sound receiver, a signal analyzer, a
measurement device, and a communicator wherein: the sound receiver
comprises a transducer converting sound energy to electrical
signals; the signal analyzer comprises a processor configured to
analyze the signals, and to identify therein a spoken command of
type 1 having a single period of voiced sound, and to activate the
measurement device responsive to the type 1 command, thereby
producing a measured value; the signal analyzer is configured to
recognize a spoken command of type 2 having two periods of voiced
sound separated by a period of substantially less voiced sound, and
to reset the measured value responsive to the type 2 command; and
the communicator is configured to communicate the measured value by
visually displaying or acoustically announcing or electronically
transmitting the numerical value.
2. The system of claim 1 wherein the sound receiver includes
amplification means with a gain of about 40 dB to about 100 dB, and
bandpass filtering means having a low-frequency cutoff between
about 10 Hz and 100 Hz and a high-frequency cutoff between about
500 Hz and 1500 Hz.
3. The system of claim 1 wherein resetting the measured value
comprises setting the measured value to zero.
4. The system of claim 1 wherein the measurement device is an
electronic counter configured to increment when activated
responsive to a type 1 command, the measured value thus comprising
the number of type 1 commands.
5. The system of claim 1 wherein the measurement device is an
electronic counter configured to count pulses from a clock
generator, and to alternately start and stop such counting
responsive to type 1 commands, thereby measuring a time interval
between successive type 1 commands, which time interval comprises
the measured value.
6. The system of claim 1 wherein the measurement device is
configured to measure one of: a distance, or a weight, or a
voltage, or a temperature, or a light intensity, or a field
value.
7. The system of claim 1 which enables a readiness indicator when
the system is ready to receive a spoken command, and which disables
the readiness indicator when the system is not ready to receive a
spoken command.
8. The system of claim 1 which further includes a running mode and
a holding mode, the system being configured to receive and respond
to spoken commands when in running mode, and to ignore all spoken
commands when in holding mode.
9. The system of claim 1 which is further configured to ignore any
subsequent type 1 commands, after performing the measurement, until
reset by a type 2 command, thereby holding the measured value
constant until so reset.
10. The system of claim 1 which further includes a multifunction
switch comprising a momentary-type pushbutton switch configured to
perform multiple functions wherein: if the switch is pressed while
power to the system is off, the switch turns on the power; if the
switch is pressed for a time interval of T1 or less, while the
power is on, the switch causes the system to alternately suspend
operation and resume operation; if the switch is pressed for a time
interval greater than T1 but less than T2, while the power is on,
the switch causes the system to reset the measured value; if the
switch is pressed for a time interval greater than T2 but less than
T3, while the power is on, the switch causes the power to turn
off.
11. The system of claim 10 wherein T1 is less than about 200
milliseconds, T2 is greater than T1 and less than about 1 second,
and T3 is greater than T2 and less than about 5 seconds.
12. A method for performing a measurement and resetting a
measurement result responsive to voice commands, the method
comprising the steps: waiting until no sound is detected for a
predetermined period Ts; then, waiting until a first sound is
detected, thereby indicating that a voice command is detected;
then, performing the measurement, thereby producing a measured
value, and communicating the measured value by means of a visual
display or an acoustical message or an electronic signal; and
waiting until no sound is detected during a period Ta, thereby
indicating that the first sound is complete; then, waiting for a
period Tg, the command being of type 1 if no sound is detected
during Tg, and the command being of type 2 if a second sound is
detected during Tg; and resetting the measured value responsive to
a command of type 2.
13. The method of claim 12 wherein Ts is about 50 to 500
milliseconds, Ta is about 10 to 100 milliseconds, and Tg is about
50 to 500 milliseconds.
14. The method of claim 12 which includes, following each command
of type 1, holding the measured value constant thereafter until
being reset by a command of type 2 or by a button press.
15. The method of claim 12 wherein resetting the measured value
comprises setting the measured value to zero.
16. The method of claim 12 which further includes producing a first
confirmation indicator following a type 1 command, and producing a
second confirmation indicator, distinct from the first confirmation
indicator, following a command of type 2, the two confirmation
indicators thereby indicating the type of voice command
detected.
17. The method of claim 12 which further includes enabling a
readiness indicator that indicates when voice commands may be
received, the readiness indicator being enabled after the period Ts
expires with no sound detected therein, and the readiness indicator
being disabled whenever any sound is detected above threshold.
18. The method of claim 12 which further includes comparing the
signals to two threshold values V+ and V-, with V+ being more
positive than V-; and a sound being detected when the signals are
greater than V+ or less than V-; and no sound being detected when
the signals remain between V+ and V-.
19. The method of claim 12 wherein performing the measurement
comprises one of: incrementing a counter, the measured value being
a count total equal to the number of voice commands detected; or
starting and stopping a counter in alternation, the counter being
configured to count pulses from an oscillator, thereby producing a
measured value proportional to a time interval between successive
voice commands; or performing a distance measurement or a voltage
measurement or a temperature measurement or a light intensity
measurement or a field measurement at a particular time determined
by the voice command.
Description
BACKGROUND OF THE INVENTION
The invention relates to voice-activated systems, and particularly
to instruments for making measurements under voice command.
Voice-activated and sound-operated devices are becoming
increasingly useful, in applications ranging from lamp dimmers to
television remote controls to surgical tools. Some products respond
to sound levels, without interpreting the sound as speech. Such
non-word-recognizing sound-activated systems include voice
recorders, games, and security systems. Other products actually
interpret sounds as spoken commands, and respond differently to
each recognized command. Voice-controlled phone dialers, GPS
systems, alarm clocks, and certain robotic toys perform such word
recognition. Even coffee pots now accept a voice command.
Practical data measurement systems, on the other hand, have not
kept pace with voice-activation technology. Event counting and
interval timing are just two types of measurements that would
greatly benefit from a hands-free command capability. Consider a
laboratory technician or a researcher handling racks of multiple
samples, counting out aliquots of reagents. The worker is
constantly counting, necessarily dividing attention between the
samples and the tally sheet. A voice-activated counting device
would be extremely beneficial. Likewise a retail store must count
stock periodically, or an assembly-line worker who has to keep
track of how many bolts he has installed on a chassis, or a cook
counting out spoonfuls, and innumerable other applications needing
a voice-activated counter. Tragically, no such device exists.
A similar deficit is apparent in the area of voice-activated
interval timing. For example in sports training, it would be very
useful if a time interval could be measured hands-free, using only
voice commands. In electronics development and testing, in many
areas of research, in psychological testing, and innumerable other
situations it would be valuable to measure a time interval by
voice. While there are many voice-activated alarm clocks on the
market, and some are marketed as "timers", in fact none of them has
a stopwatch capability with voice-controlled start and stop
function.
Many other measurement instruments would likewise benefit from
voice control. Consider a worker measuring a size or distance with
a caliper or electronic tape measure, but having difficulty reading
the device due to darkness, position, etc. A voice trigger would
solve this problem if it locked the measurement on command, to be
read later. For weighing, a voice-triggered scale would allow the
operator to focus on the load rather than the instrument. For
electrical measurements (voltage, current, resistance, frequency,
etc.) a measurement could be triggered by voice at the exact moment
desired. A voice-triggered digital thermometer would allow the user
to lock and hold the temperature reading, when holding the probe
upon a specific location. For magnetic field measurements, a
voice-activated Gaussmeter would simplify field mapping by allowing
the user to lock the measurement at will.
Most voice-activated systems under development today attempt to
interpret the meaning of a command. However, word-recognition
systems are complex and greatly increase the cost of the
application. Usually they require special "training" sessions to
calibrate recognition parameters, special microphones and software,
and still they make numerous errors. For measurement applications,
full word recognition is not necessary because only a couple of
basic command functions are needed, for example, "Go" and "Reset".
For all these applications, full word recognition is an unnecessary
expense.
What is needed is a voice-activated measuring instrument that
performs the measurement on spoken command, and resets on spoken
command. Preferably the instrument could use an inexpensive and
versatile, speaker-independent sound-interpretation strategy rather
than word recognition. Such an instrument would have innumerable
applications in research, business, sports, and home life.
BRIEF SUMMARY OF THE INVENTION
The invention is a voice-activated system and method to perform a
measurement when triggered, and to reset when triggered, by
distinct voice commands. The invention recognizes spoken commands
as single-syllable (type 1) or double-syllable (type 2) commands.
When triggered by a type 1 command, the invention performs the
measurement and communicates the measured value. Upon a type 2
command, the invention resets the measured value to zero. The
inventive system comprises a sound receiver to convert spoken
sounds to electronic signals, a signal analyzer to detect the voice
commands and recognize the command type, a measurement device that
performs the measurement responsive to type 1 commands, and a
communicator to communicate or display the measured value.
The measurement performed by the invention is any determination of
a numerical quantity when triggered by the voice command, thereby
producing a measured value. The measurement includes any
calculations needed to derive the measured value. Examples of the
inventive measurement include a tally count, such as the number of
voice commands received; a time interval, such as the time interval
between successive voice commands; a distance such as a size of an
object or a distance between objects; an electronic quantity such
as a voltage or current or resistance or frequency or amplitude; a
weight or a temperature or an intensity or a field. In each case,
the measured value is a numerical value representing the measured
quantity at the moment it is measured responsive to the voice
command.
The inventive measurement device is any means for performing the
selected measurement when triggered by the voice command. When the
measurement comprises event counting, the measurement device may be
a register in a microcontroller. For interval timing, the
measurement device may be a gated clock counter; for length
measuring, a caliper with an electronic readout; for weighing, an
electronic scale; and likewise for voltage or current or resistance
or other electronic parameter measuring, temperature or magnetic
field or light level or other environmental parameter measuring, or
any other evaluation of a numerical quantity to be made at a
specific moment upon a spoken command.
The inventive sound receiver is any transducer that converts the
sound of a spoken command into electronic signals. The sound
receiver comprises microphones, amplifiers, rectifiers, analog
electronic filters, and digital circuits.
The inventive signal analyzer is a circuit that analyzes signals
from the sound processor, and recognizes voice commands as single-
and double-syllable commands, and then triggers or resets the
measurement device. Typically the signal analyzer detects voice
commands by comparing the amplified signals to a threshold value,
and detects or recognizes a voice command when the signals exceed
the threshold, and registers that no sound is present when the
signals remain below the threshold. For bipolar signal waveforms,
the signal analyzer may compare the signals to two threshold values
V+ and V-, recognizing a sound as a command when the signals exceed
V+ or go more negative than V-.
The inventive communicator is any means for communicating the
measurement result to a user. The communicator may comprise a
visual display (such as an LED or LCD display), or an acoustical
transmitter (such as a speaker with tones or computer-generated
speech), or means for transferring the result electronically to an
external system for storage or display or further processing.
The invention recognizes a type 1 command as any single-syllable
sound comprising a single brief period of vocalization, such as
"Count", "Go", "Stop", and "Lock". The invention recognizes a type
2 command is any two-syllable sound having two voiced periods
separated by a brief non-voiced interval, such as "Reset" or
"Backup". The signal analyzer treats voiced sounds, and
particularly vowel sounds, as command sounds, while any non-voiced
speech is treated as silence or background noise. For example the
"s" in "Reset" is non-voiced and registers as a brief gap of
relative silence between the two voiced syllables in the word. Thus
type 2 commands are distinguished by the gap between two voiced
sound periods, whereas type 1 commands have a single voiced
period.
Many words have multiple syllables. The signal analyzer may
recognize a word with three syllables as a type 3 command, and so
forth. Or, the signal analyzer may recognize only the first and
second voiced periods as a type 2 command, ignoring any remaining
sounds. In the latter case, any spoken command with two or more
voiced periods separated by non-voiced periods is a type 2
command.
The reset operation typically comprises setting the measured value
to zero, although in rare cases the measured value may be set to
some other value indicating that the reset has occurred. A reset
may include additional operations such as restoring an index or
adjusting a threshold value or monitoring any switch changes or
transmitting a reset signal to an external system.
The invention includes means for excluding signals outside a
frequency range corresponding to voiced command sounds. Non-voiced
sounds, such as the "s" in "Reset" and the "t" in "Count",
generally have lower sound amplitude and higher frequency than
voiced sounds such as vowels. The instrument filters out high
frequency signals to prevent the non-voiced sounds from interfering
with the command interpretation. Specifically, the invention
includes a high-frequency cutoff that is lower than the frequency
range of the "t" in "Count", thereby ensuring that the "t" does not
register as a second sound pulse. The invention also filters out
low frequency signals, below the normal voiced range, coming from
drafts and vibrations. Typically the high frequency cutoff is in
the range of 500 to 1500 Hz and is accomplished using an inline
filter or a feedback capacitor across the amplifier. Typically the
low frequency cutoff is in the range of 10 to 100 Hz and is
accomplished using a series capacitor through which the signals
must pass.
In one aspect, the invention is a voice-activated tally counter,
and the measured value is the total count of type 1 voice commands.
Upon each type 1 command, the system increments the measured value
by one count; and upon each type 2 command, it resets the measured
value to zero. The measurement device may be an electronic counter
or memory element, which is incremented when the signal analyzer
detects a single-syllable command, and which is reset to zero when
the signal analyzer detects a double-syllable command.
In one aspect, the invention is a voice-activated interval timer or
electronic stopwatch, and the measured value is the time between
successive commands of type 1. The time measurement may be
performed by gating a counter that counts clock pulses. Clock
pulses are electronic pulses from any oscillator or pulse generator
producing periodic countable pulses. The circuit may be configured
to turn the counter on and off, or to open and close a gate on the
clock pulses, upon successive type 1 commands. In either case, the
count total is proportional to the time between the commands. The
counter may include a scaler so that the displayed value is
directly readable in seconds or milliseconds for example. The
invention may be configured so that after a time measurement,
additional type 1 commands could again start the clock counter, in
which case the displayed result is the accumulated time of all the
recorded intervals. Or, the invention may ignore any subsequent
sounds, thereby freezing the data until reset by a type 2 command.
Preferably a type 2 command also stops the counter from counting,
if it has started counting.
In one aspect, the invention measures a size or length or distance
when triggered by a type 1 command. Examples of voice-triggered
distance measuring devices include micrometers and calipers and
tape measures with electronic readouts. Typically a user measures
the size of an item by holding the measurement device against the
item and reading a result. Often it is difficult to read while
holding the device in place, due to low light level or an awkward
position for instance. But if the user removes the device from the
item for easier reading, often the value is disturbed due to the
additional motion, and gives a wrong reading. The invention solves
this problem by freezing the measurement at a precise moment of the
user's choice. The user can then remove the device and read the
measured value at leisure, without disturbing the information.
Also, it is easy for the user to abort a measurement and start
over, simply by speaking a type 2 command to force a reset, and
then another type 1 command when ready to measure.
In one aspect, the invention measures a weight when triggered by a
type 1 command, and holds the value thereafter until reset.
Laboratory scales, postal scales, metering scales, and many other
weight measuring devices are included in this category. The
invention allows the user to focus on the item being weighed
instead of memorizing the result.
In one aspect, the invention measures an electrical parameter such
as voltage, current, resistance, frequency, amplitude, and the
like. The instrument records the measurement only upon a type 1
command, and resets on a type 2 command. Multimeters and other
electronic test meters are in this category.
The invention includes a specific protocol to deal with additional
commands occurring after a measurement has been triggered. The
invention may respond to further type 1 commands by adding to the
measured value. A tally counter, for example, continues
accumulating the total count upon each type 1 command. Or, the
invention may respond to subsequent type 1 commands by replacing
the earlier result with the later result, thereby updating the
measured value upon each type 1 command. A voice-activated
voltmeter would be of that type. Or, the invention may freeze the
first measured value, ignoring further type 1 commands until being
reset by a type 2 command (or by a button press). A voice-triggered
caliper may use that feature, thereby enabling easier readout.
The invention includes two operational modes, a running mode and a
holding mode. In the running mode, triggering of the measurement
device is enabled and the system performs the measurement upon a
type 1 command and resets upon a type 2 command. In the holding
mode, all voice commands are ignored, and the measured value is
held unchanged. When the communicator is a visual display,
preferably the display continues to show the measured value while
in holding mode, but changes the display in a way easily
recognizable by the user. For example, in holding mode, the display
digits may flash on and off indicating that measurement operations
are suspended.
The invention includes means (such as a switch) for turning the
device on and off. The invention includes means (such as another
switch) for suspending and resuming measurement operations. The
invention includes means (yet another switch) for resetting the
measured value to zero manually. Alternatively, and preferably, the
invention includes a single multifunction switch that combines all
the operations of an on-off switch, a run-hold switch, and a
manual-reset switch. The multifunction switch is a momentary-type
pushbutton switch that, when pressed, causes the instrument to
respond differently depending on the state of the instrument and
the duration of the press. For example, when the instrument is off,
a press of the multifunction switch turns on the power. When the
power is on, a brief press of the multifunction switch causes the
instrument to toggle between running and holding modes. A longer
press triggers a manual reset. An even longer press turns the power
off. Such a multifunction switch is convenient and appealing to
users.
The invention includes both manual and automatic means for
adjusting the acoustical sensitivity. The sensitivity adjustment
could be increased to accommodate a user who speaks softly, or
decreased to negate a noisy environment. Examples of manual
sensitivity adjusting means include a continuously-variable
potentiometer and a multi-position click-stop slide switch, wired
to attenuate the microphone signal or to modulate the amplifier
resistors or to vary a DC level which a microcontroller then
interprets as a threshold adjustment. The automatic adjusting means
responds to changing background noise levels by monitoring the
background noise and adjusting a signal analysis parameter to
accommodate the observed level of noise. Preferably the instrument
prevents the sound of a command from unduly influencing the
adjustment, for example by suspending or retarding the automatic
adjustment whenever a voice command is detected. Preferably, the
threshold value is not reduced below a predetermined minimum value,
even in complete silence.
Since command signals include both plus and minus variations about
a mean value, the instrument should react to either positive or
negative variations. The invention may rectify the waves, thus
rendering all the signal variations positive. The rectified signals
may then be integrated or smoothed. Alternatively, the invention
may provide both positive and negative thresholds such that sound
pulses exceeding the positive threshold, or extending more negative
than the negative threshold, are interpreted as a command. Signals
remaining between the negative and positive thresholds are
interpreted as relative silence.
The invention includes detachable manual-controller means for
controlling the measurement operations manually. The controller
includes two switches or other manual means for generating signals
to control the instrument, one switch simulating a type 1 command
and the other simulating a type 2 command. Such a manual control
capability may be useful in high noise environments because it
would eliminate all background triggering. It would also be useful
in environments where the measurement has to be made silently.
Preferably, the controller disables the instrument's internal
microphone to avoid any background noise triggering the unit. The
controller may be connected to the instrument by a cable, or it may
communicate commands to the instrument wirelessly using radio or
infrared signals for example.
The invention includes a command validation indicator showing when
a command has been received, and preferably also showing which type
of command. This assures the operator that each command has been
properly heard. The indicator may be a visual indicating means such
as an LED which is flashed on, or an acoustical indicating means
such as the speaker of a headset, in which a tone is produced each
time the instrument recognizes a valid command. The indicating
means may be different when the command is a type 1 or type 2
command. For a visual indicator, a type 1 could generate a single
brief flash, while a type 2 could generate a short train of
flashes, thus showing the operator which type of command was
recognized. For an acoustical indicator, a single high tone could
indicate a type 1, while a short burst of pulses with a lower tone
indicates a type 2. In this way the operator knows that the
instrument has correctly heard each command, and also catches any
background triggers.
The invention may require a brief period of relative silence prior
to each command. For example, before accepting any command, the
instrument could monitor the background noise level and require
that the sound not exceed the threshold level for a predetermined
period, Ts, prior to accepting any command. This period of silence
ensures that previous commands are finished before accepting a new
command. The duration Ts of the period of silence must be long
enough that prior commands are finished before the system accepts a
new command, but not so long that the system appears balky.
The invention includes a specific timing protocol governing when
the selected measurement is performed responsive to a type 1
command. When the first sound is detected, there is no way to know
if the command is going to be a single-syllable or double-syllable
command. The instrument may be configured to perform the
measurement immediately, and then reset if the command turns out to
be a type 2. Or, the instrument could wait until after the first
sound is complete before performing the measurement, or it could
wait until after the entire command interpretation is complete
before performing the measurement, or it could perform the
measurement after a fixed brief delay. The favored timing strategy
depends on the application. The immediate-measurement option has
the advantage of responsiveness, but it also causes the display to
advance briefly prior to each reset, which can be jarring at first.
(This display effect occurs because the first syllable of a type 2
looks like a type 1 command.) For time interval measuring, the
fixed delay option is desirable because it avoids the unwanted
display advance, while maintaining complete time interval precision
(since the same fixed delay is added to both Start and Stop). For
distance-measuring applications, it is usually desirable to perform
the measurement immediately, since any delay could result in a
wrong result due to the device moving during the delay time. For
applications involving communication with another computer, it may
be best to perform the measurement immediately but wait until the
command is complete before communicating the result, so as to avoid
confusing the downstream device.
The invention includes signal processing means for distinguishing
type 1 commands and type 2 commands based on the temporal
distribution of sounds and silent periods. To interpret a spoken
command, the invention determines when the first sound pulse is
finished and then detects the second sound pulse, if any. The
invention determines when the first sound pulse is finished by
waiting for a time period Ta of relative silence. If further sound
exceeding the threshold is detected therein, the instrument starts
the Ta period over. The instrument continues in this fashion until
the Ta period expires with no further sound detected. At that
point, the first sound is finished. Ta must be long enough to catch
all the sound waves of the first syllable, but short enough that it
does not also collect the sound of a second syllable. Errors in
this process will make a type 1 command register as a type 2
command, and vice versa.
The invention then looks for a second sound pulse within a specific
time interval Tg which starts at the end of the Ta period. If no
further sound is detected during the Tg period, then the command is
a type 1 command. If a second sound pulse is detected anytime
during the Tg period, then the command is a type 2. Tg must be long
enough to accommodate someone speaking slowly, but not so long that
the system readiness is delayed.
The invention includes means for indicating to the operator that it
is ready to receive a verbal command. Such a readiness indicator
may be an LED for example. While the instrument is processing a
command, or waiting for the pre-command period of silence, the
readiness indicator is off. When the instrument is in holding mode
or performing a reset operation, or whenever any sound is detected,
the readiness indicator remains off. The readiness indicator turns
on only in running mode, after the silent period is met, before any
further sound is detected, and the instrument is ready to receive a
verbal command.
The invention includes powering means, which may be a battery or a
detachable power supply or a built-in power supply. When the
powering means is a battery, the instrument includes means for
indicating when the battery is running low, such as an LED that
flashes when the battery voltage sags. When the powering means is a
detachable power supply, preferably the battery is automatically
disconnected when the power supply is attached.
Structurally, the inventive system typically comprises a circuit
board in a case. The inventive sound receiver includes a
microphone, amplifiers, filters, and rectifiers on the circuit
board. The microphone preferably has sufficient sensitivity (at
least -30 dB) and the amplifier has sufficient gain (typically 40
to 100 dB) to permit signal analysis. The signal analyzer includes
analog or digital electronic means for comparing the signals to
threshold values and for measuring time intervals to discriminate
type 1 and 2 commands. The sound receiver and signal processor may
comprise a single board or separate circuit boards. Separate boards
are more expensive but provide improved isolation of the amplifier.
While the signal processing can be done with voltage comparators
and monostable timers, a much tidier solution is a digital
processor such as a microprocessor or microcontroller or gate
array.
The inventive measurement device may comprise a portion of the
signal processor, or it may be a separate device, depending on the
type of measurement. For a tally counting or interval timing
measurement, the measurement device is simply a register in the
microprocessor. For a distance measuring application, the
measurement is performed by a suitable distance measuring device
operationally connected to the circuit board.
The inventive communicator may be mounted on the circuit board,
such as a visual display, or it may be an external device
operationally connected to the circuit board, such as a detachable
speaker.
The invention is also a method. The purpose of the method is to
perform the measurement responsive to a voice command. The
inventive method comprises the following steps:
First, ambient sound is converted into an electrical signal. The
signal may be amplified to produce amplified sound signals, the
amplification being sufficient that the spoken command can be
analyzed, but not so high that oscillations occur. The
amplification may include only signals in the frequency band of
voiced vowels, and suppress signals with frequencies that are
higher or lower than this band. There is usually no need to flatten
the spectrum across this band, but resonances should be avoided.
The frequency selection may involve analog filtering, or the
amplification may be broadband while the unwanted frequency ranges
are excluded by analysis, for example using Fourier analysis.
Then, the amplified signals are processed to detect command sounds,
by comparing the signals to a predetermined threshold value or
values. Command sounds are detected when their signals exceed the
threshold. The signal processor may perform further tests such as
digitizing, smoothing, differentiating, frequency analyzing, and
pattern matching to further detect or identify commands.
The invention waits for a period of relative silence before
accepting any command. The silence period has a predetermined
duration Ts. If a signal exceeding the threshold occurs during this
time, then the interval measurement starts over, and this continues
until the Ts period expires with no further signals detected. At
that time the system is ready to receive a command. A readiness
indicator, if present, is turned on. The period of silence ensures
that any subsequent sound is a new command, and not part of some
previous activity. Ts is typically 50 to 500 msec.
Then, the invention waits for a first sound pulse exceeding the
threshold, which indicates that a voice command has started. When
such a sound is detected, the readiness indicator is turned off and
the selected measurement is performed. The measurement may be
performed immediately when the first sound is detected, or after a
fixed delay, or at the completion of the first sound. The
particular measurement steps depend on the type of measurement to
be performed. For example if the measurement is event counting,
then typically an electronic counter is incremented; and if the
measurement is a time interval between commands, then a clock
counter is started or stopped upon the first sound.
Then, the invention seeks the end of the first command sound by
waiting for a time interval Ta during which no further sound is
detected above thresholds. Additional sounds during that time are
considered part of the first syllable, and the Ta interval is
started over. When the Ta period finally expires with no further
sound detected, then the command syllable is considered finished.
Ta must be long enough to catch all of the sound of a
single-syllable command, but not so long as to include both sounds
of a double-syllable command. Typically Ta is about 10 to 100
msec.
Then, following the Ta period, the invention seeks a second sound
pulse during a time interval Tg. If the Tg interval expires with no
further sound detected, then the command was a type 1. If any sound
is detected above threshold during Tg, then the command has a
second syllable and thus is a type 2 command. When a second sound
is detected, the system performs the response function of a type 2
command such as resetting the measured value to zero. Tg must be
long enough to catch the second syllable in a slowly-spoken reset
command, but short enough that the system returns to a ready state
without much delay. Typically Tg is in the range of 50 to 500
msec.
Then, a command validation indicator (if present) such as a LED
flash or a tone is generated, and the measured value is
communicated to a display or other communication device. The
display may be updated as soon as the measured value is obtained,
or the display may be updated only after the end of the Tg period.
For a visual display, the prompt update version is more responsive.
If the measured value is communicated by computer-generated speech,
or transmitted to another device, or otherwise processed, it may be
better to delay transmission until after the Tg time so that type 1
and type 2 commands are clearly separated.
Then, the invention loops back to the step of waiting for a period
of silence Ts before receiving the next command.
The invention provides numerous valuable advantages. It allows a
user to perform tedious and distracting measurements easily and
hands-free, such as counting and resetting the count. It enables
fast time interval measurements, and arguably provides greater
precision than any manually controlled measurement. It allows the
user to measure a size or distance at a moment of choice, and then
read the result later without worrying about the measurement
changing. It allows a worker to hold multiple probes on a test
circuit while triggering a measurement hands-free. It helps control
a weighing process by allowing the user to manage the load rather
than focusing on the readout. It greatly simplifies
position-dependent measurements such as temperature and light level
and field measurements, by allowing the user to register the
measurement value at will. Most importantly, it provides all these
capabilities using a quick and simple command recognition protocol,
which is easily performed by all users without training, and is
implemented with extremely low-cost components.
BRIEF DESCRIPTION OF THE FIGURES
FIG. 1 is a sketch, partly cut away, of an embodiment of the
invention as a portable counter or timer instrument.
FIG. 2 is a sketch of size-measuring embodiments as a caliper (FIG.
2a) and electronic tape measure (FIG. 2b).
FIG. 3 is a sketch of a weighing scale embodiment as a laboratory
scale (FIG. 3a) and as an industrial conveyor scale (FIG. 3b).
FIG. 4 is a sketch of an inventive multimeter (FIG. 4a) and a
Gaussmeter (FIG. 4b).
FIG. 5 is a chart showing the sequence of intervals in a command
response.
FIG. 6 is a flowchart showing the steps of the inventive
method.
DETAILED DESCRIPTION OF INVENTION
Referring to FIG. 1, an embodiment according to the invention is an
instrument comprising a case 101 (shown partly cut away), a
microphone 102, a circuit board 103, and a display 104. The case
101 is a plastic enclosure with an aperture for the display 104,
plus a battery 105, a multifunction switch 106, a gain switch 107,
a mode switch 108, a communications connector 109 for a detachable
microphone (not shown), and a power connector 110 for a detachable
power supply (not shown).
The microphone 102 is a small electret transducer such as the
CMC-5044PF-A, with sufficient sensitivity (preferably at least -30
dB) in the vocal frequency range to detect spoken commands, and is
wired to generate electrical signals related to the command
sounds.
The circuit board 103 includes an amplifier 111 which is configured
to amplify the signals from the microphone 102, preferably
excluding signals having frequencies outside the voiced frequency
range. The amplifier 111 may be an LM358 having two op-amps wired
as inverting, AC-coupled gain stages in series. The circuit board
103 includes a microcontroller 112 which receives the amplified
signals and is programmed to identify the voice commands as type 1
or type 2 based on their sound pulses. Preferably the
microcontroller 112 has an ADC capability to digitize the signals.
The microcontroller also performs as the measurement device by
incrementing a counter within the microcontroller 112, or gating a
timer within the microcontroller 112, when so commanded. The
microcontroller 112 is also configured to reset the measured value
to zero upon a double-syllable voice command. The microcontroller
112 may be a PIC16F690 with an 8.0 MHz crystal clock.
The display 104 is a 4-digit 7-segment common-cathode LED display
such as the LTE-4727JR. The microcontroller 112 drives the display
104 by illuminating the segments of one digit at a time, rapidly
circulating among the four digits to produce a visually continuous
result. The embodiment also provides a low-battery indicator
comprising one of the decimal points in the display 104. The
microcontroller 112 monitors a voltage line and detects when the
power sags, and then illuminates or flashes that decimal point to
alert the user that the battery 105 is low. The embodiment also
provides a readiness indicator comprising another decimal point in
the display 104. When the system is ready to receive a spoken
command, following a requisite period of silence, the
microcontroller 112 illuminates the readiness indicator, and turns
it off as soon as the first command sound is detected. In this way
the user knows when the system is ready for a voice command.
The multifunction switch 106 is a tactile switch (a momentary-type
pushbutton switch with a snap action that can be felt when pressed)
wired to turn the power on when pressed. The microcontroller 112
monitors voltage from the multifunction switch 106 and is
programmed to measure the duration of the press and to respond to
different press durations accordingly. When the system is running,
a brief tap of the multifunction switch enables or disables the
measurement function by alternating between the running mode and
holding mode. Upon a longer press, lasting about 0.5 seconds, the
microcontroller 112 resets the measured value to zero. For an even
longer press, lasting about 1.5 seconds, the microcontroller 112
causes the system to power-down.
The gain switch 107 is a 3-position slide switch connected to three
attenuators (not shown) and to the microphone 102 so that the
signal from the microphone 102 must pass through the attenuator
that is selected by the gain switch 107. The gain switch 107 may be
mounted on the side as shown, or anywhere of convenience on the
case 101. In another embodiment, the gain switch 107 selects either
the feedback resistor or the gain-set resistor in the amplifier
111, which accomplishes the same sensitivity variation as the
attenuation version although it requires one additional
interconnect wire.
The mode switch 108 is a toggle switch connected to the
microcontroller 112 to select the type of measurement that is
performed. The embodiment shown performs two measurements, a
counter measurement and a timer measurement, as selected by the
mode switch 108. The mode switch 108 may be mounted on the side, as
shown, or anywhere of convenience on the case 101. In different
embodiments, the selecting may occur only upon power-up, or the
selecting may occur only during a reset operation, or the selecting
may occur at any time the user changes the mode switch 108.
The communications connector 109 is a 2.5 mm, 3-conductor, stereo
jack which receives a mating plug attached to an external
peripheral (not shown), such as a headset or a wireless microphone
receiver or a manual controller. The communications connector 109
is wired to the circuit board 103 to convey signals between the
circuit board 103 and the peripheral. The communications connector
109 also includes an internal switch (not shown) that disconnects
the internal microphone 102 from the circuit board 103 whenever the
peripheral is plugged into the communications connector 109, so
that the peripheral can control the operation without interference
from the internal microphone 102. The communications connector 109
may be mounted on the front, as shown, or elsewhere. The microphone
102 may be mounted on the circuit board 103, as shown, or adjacent
to the communications connector 109. Mounting the microphone 102
close to the communications connector 109 usually increases the
distance from the microphone 102 to from the amplifier 111, but
minimizes the number of interconnects needed.
The power connector 110 is a 2.1 mm, 2-conductor, center-positive
coaxial male plug onto which an external 9-volt DC power supply
(not shown) can be connected. The power connector 110 includes an
internal switch (not shown) that disconnects the battery 105 when
the power supply is plugged in. The power connector 110 may be
mounted on the front, as shown, or elsewhere on the case 101.
FIG. 2 shows inventive embodiments configured to measure a
distance. In FIG. 2a, an electronic caliper 201 measures the size
of any item placed between the jaws 202 (shown in the closed
position). A microphone 203 receives voice commands. The caliper
201 performs the size measurement continuously, and display it
continuously in the display 204, until triggered by a type 1
command, at which point the display 204 is frozen until reset. The
result can be zeroed, and the system readied for another
measurement, by a type 2 command or using the button 205. The user
can easily read the measurement by issuing a type 1 command, and
then removing the item from the jaws 202, since the result in the
display 204 remains unchanged until reset.
FIG. 2b shows an electronic tape measure 210 configured to measure
a distance using the retractable tape 211 and an internal
optical-encoder wheel (not shown). A microphone 212 accepts voice
commands to trigger the measurement, and a display 213 shows the
result. A button 214 toggles between holding mode and running
mode.
FIG. 3 shows inventive embodiments configured for weighing
something. In FIG. 3a, a laboratory scale 301 measures the weight
of any item placed thereon and displays the result in the screen
302. A switch 303 causes the scale to "tare" or readjust its zero
point, an operation commonly performed before every weighing. A
microphone 304 receives voice commands and is configured to trigger
a tare adjustment upon each type 2 command, and to lock the weight
reading upon each type 1 command. Thus a user can place an empty
container on the scale 301, speak "Reset" to tare the container,
then add contents to the container, then speak "Lock" to hold the
reading which shows the weight of the contents. The user can then
remove the container at will, while the net weight remains
displayed (until reset by a button 303 press).
FIG. 3b shows an industrial weighing scale 310 with a conveyor belt
311. The conveyor belt 311 transports an object 312 to the scale
310 for weighing, and then stops. An operator checks the object
312, readjusting its position if necessary, and then speaks a type
1 command into the microphone 313. The type 1 command causes the
scale 310 to record the weight in an internal memory, and also to
generate an acoustical command-validation signal comprising
computer-generated speech using the speaker 314. The
computer-generated speech may comprise a standard response such as
"measurement completed", or it may convey the measured value as
"three point four kilograms". The operator then speaks a type 2
command into the microphone 313, which causes the conveyor belt 311
to start up again, bringing the object 312 to a printer (not
shown). The scale 310 communicates the measured value to the
printer, which prints the weight on the object 312.
FIG. 4 shows inventive embodiments configured to measure
electromagnetic quantities. In FIG. 4a, a multimeter 401 is able to
measure voltage, current, resistance, and other electronic values
when the leads 402 are contacted to a test circuit (not shown). The
particular measurement is selected by a rotatable selector switch
403. The multimeter 401 may also measure temperature using a
temperature probe (not shown). The result is displayed on the
screen 404. A microphone 405 receives voice commands and, upon each
type 1 command, locks the display 404. Upon each type 2 command,
the multimeter 401 unlocks the display 404 and resumes performing
the measurement. Thus the measured value is held for easy reading,
even after the leads 402 have been removed from the test
circuit.
FIG. 4b shows a Gaussmeter 410 which measures magnetic fields
passing through the end of a Hall-effect probe 411, and displays
the value on a screen 412. As anyone knows who has used such a
meter, the displayed value is extremely sensitive to the position
and orientation of the probe 411. Typically the operator positions
the probe in a magnetic structure, and then reads the display 412
and tries to memorize the reading before it changes. Yes, it's
difficult. Fortunately, the invention includes a microphone 413
that locks the display 412 for easy reading or recording upon each
type 1 command, and then resumes the measurement upon a type 2
command. Some prior art meters include a peak-detect option that
holds the display at the maximum value observed, but this is
frequently erroneous due to positioning errors, particularly in a
nonuniform field. The invention eliminates the problem by allowing
the operator to control exactly when the measurement is taken, by
voice command alone.
FIG. 5 is a chart showing the sequence of intervals in the
inventive command response method. At the top of the chart is a
graph of the amplified sound signal versus time. The sound plot
includes the two sound pulses comprising a type 2 command. The
positive and negative threshold values V+ and V- are shown as
dashed lines. The second trace labeled "command" indicates the
times when the signal exceeds the threshold values and is
interpreted as a spoken command. Sound is detected as a command
whenever the signal exceeds the threshold. When the signal remains
between the V+ and V- values, no sound is detected, and the signal
comprises background noise or relative silence.
The next trace in FIG. 5, labeled "ready light", shows when the
instrument is ready to receive a voice command. Initially, the
instrument is not ready because it is waiting for a silence period
Ts to complete. During the silence period, no sound exceeding the
threshold values should be detected; and if a sound is detected,
then the silence period starts over. Here Ts is 200 msec, which is
long enough to ensure that prior commands or noise excursions are
complete, but not so long that the system becomes balky.
Then, after the silence period is satisfied, the ready light turns
on indicating to the user that the system is waiting for a
command.
When the first sound pulse occurs, the ready light turns off. The
ready light remains off until the completion of the command
response plus an additional Ts period of silence.
The next trace in FIG. 5, labeled "measurement", shows when the
selected measurement is performed, such as incrementing a counter,
or gating a timer, or reading a weight or a distance value. The
selected measurement is performed immediately following the first
sound detected after the Ts silent period. Optionally, the
measurement could be performed after a fixed delay or at the
completion of the command interpretation scheme.
The next trace in FIG. 5, labeled "gap interval", shows a
predetermined period Ta following the first sound pulse, Ta being
indicated by vertical dotted lines. No additional sound should be
detected during the Ta period. If another sound occurs during the
Ta time, the interval is restarted. This continues until the Ta
time expires with no additional sounds detected. In this way the
system finds the end of the first sound pulse. Here Ta is 30 msec,
which is long enough to catch any further sound waves comprising
the first syllable, but short enough that the second sound pulse
will not be interpreted as a continuation of the first sound
pulse.
A predetermined gap period Tg then starts at the end of Ta. During
Tg, the system listens for a second sound pulse. If no second sound
pulse is detected by the end of Tg, then the command is a type 1.
If a second sound pulse is detected before the end of Tg, as shown
in the figure, then the command is type 2. Here Tg is 200 msec,
which is long enough to accommodate the second syllable of a spoken
Reset command, even if the operator speaks slowly, but short enough
that two type 1 commands are never interpreted as a type 2
command.
The next trace in FIG. 5, labeled "validator", shows the
command-validation indicator such as an LED flash or an acoustical
tone, which indicates to the user that a valid command has been
received.
After the end of the command-validation indicator pulse, plus an
additional Ts period of silence, the ready light again turns on and
the system is ready to receive the next command.
FIG. 6 is a flowchart showing the various timing loops of the
inventive method for performing voice-controlled measurements. The
instrument begins by waiting a time Ts for background noise to
settle. If any sound is detected above background during this time,
as indicated by the interrogator labeled "sound?", then the
interval is started over, continuing until a full Ts period of
silence is obtained.
Then in the box labeled "Wait for first syllable", the instrument
monitors the amplified microphone signal for any sound above
threshold. When such a sound is detected, the selected measurement
is performed. The particular measurement depends on the
application, and may involve incrementing a tally count, toggling a
timer on or off, or recording a voltage value or a weight or a
distance value for example.
Then in the box labeled "Start Ta period" the instrument seeks the
end of the first command sound by measuring a time period Ta, and
if any sound is detected during that time, it restarts the period,
continuing until a full Ta interval expires with no sound
detected.
Then the instrument seeks a second sound in an interval Tg. If a
second sound pulse is detected during that time, then the command
is a type 2 command and the system performs a reset function. If no
sound is detected during Tg, then the command is a type 1 command
and no such reset is performed.
The system then generates a command validation indicator such as an
LED flash or an acoustical tone, the indicators being different for
type 1 and type 2 commands. After the respective indicator, the
system again seeks an initializing period of silence and starts
over.
Optionally, the method includes an additional step of measuring
background noise, such as the amplitude of the amplified signals,
and adjusting the threshold higher in high noise environments, or
lower (that is, more sensitive) in quieter environments. The
adjustment may proceed continuously; or it may be interrupted
whenever a command sound is detected, thereby preventing the
command sound from unduly influencing the threshold adjustment.
The embodiments and examples provided herein illustrate the
principles of the invention and its practical application, thereby
enabling one of ordinary skill in the art to best utilize the
invention. Many other variations and modifications and other uses
will become apparent to those skilled in the art, without departing
from the scope of the invention, which is to be defined by the
following claims.
* * * * *